Between-item similarity frees up working memory resources through compression: A domain-general property

Area of Science:

• Cognitive Psychology
• Neuroscience

Background:

• Working memory (WM) relies on information compression.
• Compression benefits WM performance and can free up resources for other information.
• Traditional views suggest between-item similarity hinders WM performance.

Purpose of the Study:

• To investigate how between-item similarity affects working memory (WM) performance.
• To determine if between-item similarity can free up WM resources through compression.
• To examine this effect across semantic, phonological, visuospatial, and visual domains.

Main Methods:

• Four experiments were conducted across different domains (semantic, phonological, visuospatial, visual).
• Participants' recall performance was measured under conditions with varying levels of between-item similarity.
• A condition minimizing whole-list similarity was used as a baseline for comparison.

Main Results:

• Between-item similarity enhanced recall for similar items.
• Between-item similarity also improved recall for less compressible items within the same list.
• This enhancement was proactive, not retroactive, and occurred across all tested domains.

Conclusions:

• Between-item similarity can be leveraged to free up WM resources via compression.
• This compression mechanism allows for efficient reallocation of WM resources.
• The findings challenge traditional views and highlight a novel mechanism for WM enhancement.